Should I manually cleanup/teardown in Cloudflare Workers?

[g-otn]

Hello!

I'm using postgres inside a Cloudflare Worker.

I'm wondering: Assuming all operations are properly awaited, if it's necessary to manually call sql.end() at the end when using Cloudflare Workers?

This library's docs says Cloudflare Workers allows this package to use the TCP socket API, which seems to automatically cleanup connection "on EOF". This, alongside the following examples, seems to imply it's not necessary:

• https://github.com/porsager/postgres?tab=readme-ov-file#cloudflare-workers-support
• https://developers.cloudflare.com/workers/tutorials/postgres/#5-interact-with-the-products-database

However, I found this other example using this same library where the sql.end is observed:

• https://developers.cloudflare.com/hyperdrive/examples/connect-to-postgres/supabase/#3-use-hyperdrive-from-your-worker

Is this something specific to Hyperdrive connections, are some of the docs outdated or am I missing something else?

Sorry if it's not the right place to ask, but since the docs mention Cloudflare Workers with an example, I thought I'd ask.